LotuS run incomplete in sdm step

[qinglong wu]

Dear Falk,

I installed LotuS in my linux box. But when I test the 454 sample data set from LotuS web source, I can not go through the whole analysis, stopped by demultiplux step, no UPARSE clustering. I have installed the UPARSE and configure it the the lOTUs.cfg file.

I pasted the LotuS_runlog info. for your identification as below:

_________________________________________________________________

[qinglonw@tcmc_sandbox2 ~]$ perl /export/apps/lotus_pipeline/lotus.pl -i ./454data/ -map ./454data/map.txt -s ./454data/sdm_454.txt -p 454 -o ./454data/output -refDB SLV -saveDemultiplex 1 -c /export/apps/lotus_pipeline/lOTUs.cfg

=========================================================================

          LotuS 1.561

=========================================================================

COMMAND

perl /export/apps/lotus_pipeline/lotus.pl -i ./454data/ -map ./454data/map.txt -s ./454data/sdm_454.txt -p 454 -o ./454data/output -refDB SLV -saveDemultiplex 1 -c /export/apps/lotus_pipeline/lOTUs.cfg

RefDB SLV requested. Setting similarity based search to default Blast option to search SLV.

SLV

Using Silva SSU ref seq database.

Checking for updates..  Your LotuS version is up-to-date!

=========================================================================

          Reading mapping file

=========================================================================

Sequence files are indicated in mapping file.

Running UPARSE de novo sequence clustering..

Running fast LotuS mode..

------------ I/O configuration --------------

Input=   ./454data/

Output=  ./454data/output

TempDir= ./454data/output/tmpFiles/

------------ Configuration LotuS --------------

Sequencing platform=454

AmpliconType=SSU

OTU id=0.97

min unique read abundance=2

UCHIME_REFDB, ABSKEW=/export/apps/lotus_pipeline//DB//rdp_gold.fa, 2

OTU, Chimera prefix=OTU_, CHIMERA_

Similarity search with Blast

ReferenceDatabase=SILVA

RefDB location=/export/apps/lotus_pipeline//DB//SLV_123_SSU.fasta

TaxonomicGroup=bacteria

--------------------------------------------

Demultiplexed input files into single samples, no quality filtering done

=========================================================================

          Demultiplexing input files

           elapsed time: 2 s

=========================================================================

This is sdm (simple demultiplexer) 1.35 beta.

NO filtering will be done on your reads (just rewriting / log files created).

Writing demultiplexed files to: ./454data/output/demultiplexed/

Reading Fasta + Quality file.

fasta file : ./454data//Anh.1.fna

quality file : ./454data//Anh.1.qual

   [#############################################################] 100.00%

sdm 1.35 beta

Input File:  ./454data//Anh.1.fna

Output File: ./454data/output/tmpFiles//demulti.fna

No valid Filter file provided; no filtering done on files

Reading Fasta + Quality file.

fasta file : ./454data//Anh.2.fna

quality file : ./454data//Anh.2.qual

   [#############################################################] 100.00%

sdm 1.35 beta

Input File:  ./454data//Anh.2.fna

Output File: ./454data/output/tmpFiles//demulti.fna

No valid Filter file provided; no filtering done on files

Evaluating and writing dereplicated reads..

Dereplication:

Accepted 19,616 unique sequences ( 2 ); average size in this set is 5.0808.

Uniques with insufficient abundance: 282,803 not passing derep conditions

Time taken: 17.58s

./454data/output/demultiplexed/

Finished task

_________________________________________

Hope to receive your comments soon.

Thanks a lot.

Qinglong from Baylor College of Medicine

[Falk Hildebrand]

Dear Qinglong,

you choose the option "-saveDemultiplex 1". This *only* demultiplexes files and stops the pipeline afterwards, the mode is not really thought for working with data, it is more thought for publishing the data, which is why no quality filtering is done. You can use -saveDemultiplex 2, which will demultiplex your data into quality filtered single files and continue the pipeline afterwards, but if you just want to play with your data, try different clustering algorithms etc, I recommend to use -saveDemultiplex 0, as this saves on I/O and makes the pipeline in general faster.

best,
Falk

[qinglong wu]

Dear Falk,

Thanks for getting back to me. Got your explanation now. BTW, there is no "-saveDemultiplex 2" option in the lotus.perl command with "--help" option, but that option is available in wen documentation.

I meet another problem with the "-simBasedTaxo" option:

when the default DB (greengenes) was used, RDP classifier (default) and BLAST (need specified in command) can be used for taxonomy assignment; when I chose SILVA database, the BLAST was automatically used even though I specified the assign method with RDP classifier (-simBasedTaxo 0).

I attached part of the LotuSlog file for your information.

=========================================================================

          Removed 28 chimeric

          and 0 phiX contaminant

          OTUs (2105 read counts) from abundance matrix,

          282 OTUs remaining.

=========================================================================

cp: cannot stat ‘.output_SILVA_RDP//tmpFiles//RDPotus.tax’: No such file or directory

=========================================================================

          Assigning taxonomy against reference using BLAST

          elapsed time: 14 s

=========================================================================

Least common ancestor (LCA) assignments ver 0.16                alpha

Loading tax DB..

Look forward to receiving your comments.

Thanks.

Qinglong

[Falk Hildebrand]

Dear Qinglong,

Can you sent me your entire lotus command? Most likely a combination of parameters that is triggering some unwanted behabiour.

Will fix the saveDemulti help in next version, thanks.

Best,

Falk 

[qinglong wu]

Dear Falk,

I have attached my commands for your overview:

>perl /export/apps/lotus_pipeline/lotus.pl -i ./test_2/input_reads/ -map ./test_2/CDGABA.LOTUS_map.tsv -s ./test_2/sdm_miSeq.txt -p miSeq -o ./test_2/output_SILVA_RDP/ -c /export/apps/lotus_pipeline/lOTUs.cfg

Results in log file is "taxonomy assigned by RDP using GG DB".

>perl /export/apps/lotus_pipeline/lotus.pl -i ./test_2/input_reads/ -map ./test_2/CDGABA.LOTUS_map.tsv -s ./test_2/sdm_miSeq.txt -p miSeq -refDB SLV -o ./test_2/output_SILVA_RDP/ -c /export/apps/lotus_pipeline/lOTUs.cfg

Results in log file is "taxonomy assigned by BLAST using SILVA DB".

>perl /export/apps/lotus_pipeline/lotus.pl -i ./test_2/input_reads/ -map ./test_2/CDGABA.LOTUS_map.tsv -s ./test_2/sdm_miSeq.txt -p miSeq -simBasedTaxo 0 -refDB SLV -o ./test_2/output_SILVA_RDP/ -c /export/apps/lotus_pipeline/lOTUs.cfg

Results in log file is still "taxonomy assigned by BLAST using SILVA DB".

I also attached my lOTUs.cfg for your review.

Thanks.

Qinglong

[Falk Hildebrand]

Hey Qinglong, 

actually the -refDB option, if set, always overrides the -simbasedTaxo command (this is due to an internal logic on defaults). I'll insert a warning about this parameter in the next lotus release and actually set it to override the -refDB parameter. Your configuration file looks completely ok.

cheers,

Falk

[qinglong wu]

Thanks, Falk! 

Everything is clear to me now.

Best,

Qinglong